Since the mid-eighties both social and technical scientists
have
shown a growing interest in indigenous soil classifications.
More
and more researchers from a variety of disciplines are now
working
with such classifications. But is it actually that easy to
capture
an indigenous classification, and is there perhaps a danger of
overlooking certain complications? This article examines some
of
the issues involved and offers suggestions for further
research.
This interest in indigenous soil classifications is fully justified. Working with indigenous soil classifications offers several important benefits. For one thing, when a detailed inventory of soil resources is required, indigenous classifications are often much faster and cheaper than conventional soil survey techniques (Schutjes and Van Driel, 1994:2-3). In addition, the use of local soil terms can considerably facilitate communication between farmers, extension workers and researchers (Sikana, 1993; Warren, 1992:4). A third advantage is that a local soil taxonomy can offer important insights into the land-use considerations of farmers and the soil/plant interactions which they deal with (Tabor, 1990; Dialla, 1993; Lynham et al, 1986)
The rapid increase in the number of studies on indigenous soil taxonomies suggests not only that the importance of working with such taxonomies is now widely recognized, but also that it should be fairly easy to capture and work with this form of indigenous knowledge. In practice, however, there are a number of complications. Three of the most important of these are discussed below.
Heterogeneity of local soil knowledge
An initial complication is the heterogeneity of local soil
knowledge: even within a single village similar soils may be
referred to by different names (Tabor, 1990; Schutjes and Van
Driel, 1994). This can be of considerable importance, since
one of
the main reasons to opt for indigenous classification is in
order
to facilitate communication between farmers, extension workers
and
researchers. If a local classification has a high spatial
variability, then in effect everyone is still speaking a
different
language. In such cases, it might be more useful to adopt a
common,
more general, classification capable of integrating aspects of
the
relevant local and scientific classifications.
Correlating indigenous and scientific soil
types
Another complication is the problem of correlating local soil
types
with scientific classifications (Schutjes and Van Driel, 1994;
Sikana, 1993), which are often based on other premises than
indigenous classifications. The scientific classifications
tend to
focus on the deeper soil horizons, which represent the more
fixed
characteristics of a soil. Moreover, they strive for universal
applicability, and often make use of pedogenetic criteria for
their
semantic organization. Indigenous classifications, on the
other
hand, focus on the surface horizons that are often most
relevant
for the agricultural evaluation of a soil. They are usually
site-
and application-specific, and their semantic structure often
represents the land-use interests of the community. Another
important difference is that scientific classifications are
concerned primarily with invariant soil characteristics, which
permit the unambiguous assignment of a soil to a single class,
whereas indigenous classifications are concerned with the
direct
evaluative characteristics, which are affected by land-use
(Schutjes and Van Driel, 1994). This implies that indigenous
soil
maps are to some extent dynamic (the classification of certain
land
units will change over time) and cannot be dealt with in the
same
way as static scientific soil maps.
Insider or outsider perspective
Judging by the literature (e.g., Tabor, 1990; Sikana, 1993;
Schutjes and Van Driel, 1994), the first two problems are
quite
widely recognized. However, there is a third problem, one
which is
often overlooked. When working with indigenous knowledge, some
researchers all too easily assume that as long as they devote
attention to indigenous knowledge they will indeed develop a
true
insider or 'emic' perspective on the object of study, as
opposed to
an outsider or 'etic' perspective. This is decidedly not the
case.
It is the approach used to capture an indigenous
classification
which largely determines the degree to which one's knowledge
of the
local soil classification may be said to be that of an
insider.
Basically, there are two different approaches. The first is what might be called the anthropological approach, which relies on the use of well- defined ethnoscienctific procedures. An example of such a well-structured approach is the study by Furbee (1989), which focuses on local soil classification in Peru. She began by collecting soil names from just a few informants, by eliciting and frame interviewing. Next, she cross-checked those names with a larger number of informants, in order to do away with initial inconsistencies and redundant terms. Ten informants were then asked to perform a general sorting task with a view to constructing taxonomic trees. On the basis of this information, a questionnaire on the characteristics of the soils was constructed and completed by five other informants. Next, a number of soil samples were given to the informants to be identified, and a triadic sorting task was completed by two of the informants, in order to verify the earlier results. Finally--and this step was unique to her study--she proceeded to model local soil knowledge with the aid of an Expert System.
The second approach, which might somewhat irreverently be referred to as the 'unscientific' approach and is most widely used by the more technical scientists, makes use of individual and group interviews and field identification of soils by farmers.
The first approach is clearly more complex**1 than the second, and this is probably one of the reasons why in practice the second, unscientific, approach is more widely used. However, a more important difference is that while both approaches probably produce the same taxonomic categories (terms), they are likely to result in a very different semantic organization of those categories (in other words, they will produce a different taxonomic tree). The anthropological approach leads to an emic semantic organization of the categories, while the unscientific approach results in an etic organization of the categories. In the first case, the ethnoscienctific procedures (such as general sorting and triadic sorting) help to bring out the indigenous semantic organization of the soil terms. By relying primarily on interviewing techniques, the second approach tends--consciously or unconsciously--to produce an outsider taxonomic organization. The fact that this is more than just a theory has been illustrated by a number of case studies.
Dialla (1993) and Schutjes and Van Driel (1994) both present indigenous soil classifications of the Mossi of Burkina Faso which are based largely on interviews. Their studies were not carried out in the same region, and although this may explain some of the differences in the categories, it cannot explain the totally different semantic organizations of those categories. The study areas are not too far apart and there are no radical differences in environmental conditions. Therefore, given the fact that we are dealing with the same culture, the semantic organization may be expected to be more or less similar. Nevertheless, Dialla (1993) recognizes four major classes based on texture, whereas Schutjes and Van Driel (1994) see topographical position (or terrain type) as the major grouping, and texture as a lower-order diagnostic property. In fact, some of the terrain types distinguished by Schutjes and Van Driel (1994) are soil types in Dialla's (1993) classification. Lynham et al (1986:18) ran into this same problem when they complained that Mauritanian farmers frequently use the same names for soils and landscapes.**2 It will be clear that in such cases explicit and implicit assumptions on the part of the researchers have shaped the output. The degree to which this has taken place is, however, difficult to establish.
The above example demonstrates that it is extremely difficult to gain the emic (insider) perspective by means of interviewing techniques alone. Knowing which questions to ask, in what order, and in what form, is an art, one that even when it has been mastered by the researcher is still probably not the easiest way to get at the semantic structures underlying a classification. Warren (1992) may have been acknowledging the limitations of relying on interviewing techniques when he decided to present several taxonomic trees, rather than just one. For the Yoruba, for example, he gives one classification based on texture, one on colour, and one on fertility. For the Haussa, he gives separate classifications for agricultural and non-agricultural soils. In producing more than one taxonomic tree, he provides an impression of the diagnostic criteria without pretending to know the full indigenous semantic organization of the terms.
We should not conclude from this discussion that the anthropological approach is the only acceptable one for studying indigenous classifications. On the contrary, as long as we keep in mind the limitations of the unscientific approach, there is no objection to using it. It is not difficult to think of cases where the categories themselves are the main focus of interest and the semantic organization is of lesser importance (for example, where we are interested only in determining the distribution of different soils among households). In other cases, it may be preferable to work with the local taxa but within our own semantic organization. Only in those cases where the aim is to understand how local people see the land and soils of their community will the more complex anthropological approach be required.
The most important lesson to be learned from this third complication involved in working with indigenous soil classifications is that doing research on indigenous knowledge is not the same thing as working from the emic perspective. Depending on our research techniques, we develop either an emic or an etic perspective. It is our objectives which should determine which of the two is most desirable.
Conclusion and topics for further
research
The heterogeneity of local soil knowledge, the problem of
correlating indigenous and scientific soil types, and the
difference between the emic and the etic approach are among
the
major complications that must be kept in mind by anyone
working
with indigenous soil classifications or, for that matter, with
any
other kind of indigenous taxonomy.
There are many other issues which have to be studied if we are to deepen our understanding of indigenous soil classifications and the complications which they entail. Little is known, for example, about the degree to which the semantic organization of local soil classifications is culture-specific. What is the influence, on the one hand, of cultural characteristics (such as language and religion), and on the other hand, of environmental conditions and practical considerations (such as the particular crops grown and the available agricultural technology) on the semantic structure of local taxonomies? For example, does a change in agricultural technology lead to corresponding changes in the semantic structure? Do neighbouring cultures within the same agro-ecological zone usually have a similar semantic organization of their soil classifications? Do communities within a single culture but living in different agro-ecological zones have different semantic organizations of their classifications? These are important questions, which must be answered if we are to expand the use of indigenous classifications in research and development work.
Another interesting topic for research is the study of indigenous soil classifications in pastoral and other non-agricultural societies. There have been few, if any, studies on this subject, which might well provide interesting new ways of looking at soils. In general, a confrontation with other concepts of resource use and classification (the emic perspective) is capable of enhancing the scientific understanding of those resources and their management.
Thus working with indigenous soil classifications can make development work more effective, by providing a better understanding of the issues and options of the local farmers. It can also lead to a more thorough understanding of the local ecosystem, while offering a number of new angles from which to examine the ecological problems of the industrialized world. Given the present environmental crisis, other, more holistic, approaches to resource use and classification might well be worth considering.
David Niemeijer
Department of Irrigation and Soil and Water Conservation
Wageningen Agricultural University
Nieuwe Kanaal 11
6709 PA Wageningen
The Netherlands.
Tel: +31-8370-84190.
Fax: +31-8370-84759
E-mail: DNiemeijer@rcl.wau.nl
Acknowledgements
The author would like to thank Hector Adegbidi, Rosa Lehmann,
and
Valentina Mazzucato for their helpful comments on earlier
versions
of this manuscript.
References
Dialla, B.E. (1993) 'The Mossi indigenous soil classification
in
Burkina Faso', Indigenous Knowledge and Development
Monitor 1(3):17-18.
Furbee, L. (1989) 'A folk expert system: soil classification in the Colca Valley, Peru', Anthropological Quarterly 62(2):83-102.
Lynham, M.B., T.R. Frankenberger, W. Phelan, H. N'Gaide, P. Stone, J. Tabor and N. Harouna (1986) Farming systems research along the Senegal river valley: a rainy season reconnaissance survey in Guidimaka, Gorgol and Brakna regions. Arizona: College of Agriculture, University of Arizona.
Schutjes, A.H.M. and W.F. Van Driel (1994) 'La classification locale des terres par les Mossi: Paysan et pedologues parlent-ils le meme langage?'. Paper presented at the Premier colloque International de l'AOCASS: gestion durable de sols et de l'environnement en Afrique Tropicale. Ouagadougou.
Sikana, P. (1993) 'Mismatched models: how farmers and scientists see soils', ILEIA Newsletter 9(1):15-16.
Tabor, J. (1990) 'Ethnopedology: using indigenous knowledge to classify soils', Arid Lands Newsletter 30:28-90.
Warren, D.M. (1992) 'A preliminary analysis of indigenous soil classification and management systems in four ecozones of Nigeria.' Discussion Paper RCMD 92/1. International Institute for Tropical Agriculture (IITA) and Nigerian Institute of Social and Economic Research (NISER).
Endnotes
**1 While the anthropological approach is
more
complex, it may not be more laborious, since it does not
necessarily involve working with large numbers of
informants.
**2 This attempt to distinguish between
soils
and landscapes is typical of most scientific classifications.